Re: [Ql-Users] A feature or a bug?

[Ralf Reköndt]

No, that does not work the same way, as if you set the three SV to "". Just 
try. I always set SV_DDLST to 0.


original = peek_w (peek_l (sv_base+176)) : rem SV_DDLST
poke_w peek_l (sv_base+176) , 0 : rem set to zero, TK2 has no default now
ferr = fop_in (#5,fname$) : try to open_in the file

error testing here

Value back:
poke_w peek_l (sv_base+176) , original : back to the TK2 default, the 
devstring is still in memory


So there is nothing changed and the old devstring location is put, where it 
was before. No TK2 defaults.


Cheers...Ralf

- Original Message - 
From: "Wolf" <w...@wlenerz.com>

To: <ql-us...@q-v-d.com>
Sent: Wednesday, June 15, 2016 6:49 AM
Subject: Re: [Ql-Users] A feature or a bug?



Hi,
I often use something like this for ftest

100 DEFine FuNction ftest2(file$)
110 LOCal dad$,ded$,prd$,a
120   dad$=DATAD$
130   ded$=DESTD$
140   prd$=PROGD$
150   DATA_USE ""
160   DEST_USE ""
170   PROG_USE ""
180   a=FTEST(file$)
190   DATA_USE dad$
200   DEST_USE ded$
210   PROG_USE prd$
220   RETurn a
230 END DEFine ftest2
240 :

HTH

Wolfgang 


___
QL-Users Mailing List

[Ql-Users] A feature or a bug? No, a feature

[Wolfgang Lenerz]

Hi,

> 
> for some years I have a similar behaviour on my SMSQ/E bug list.

Ok, but it's not a bug.

> Just enter VIEW & . In case DATAD$ is set to a valid WINx_ (or FLPx_)
> path (e.g. WIN1_ or WIN1_sys_) then you get the raw content of the directory
> table listed.

Yes, that is normal, view also uses the data default.


> The RAM device driver seems not to be affected by this problem.


Yes it is, it's just that you don't have enough files in the ram disk -
copy some files into it, and you will see that it shows the same bevaviour.

HTH

Wolfgang
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Derek Stewart]

OK, sorry, but message implied this.

Back to sleep again.

Regards,

Derek

On 14/06/16 18:49, Marcel Kilgus wrote:

Derek Stewart wrote:

I am a little confused, where does DEV1_ default to DATAD$


Never. I said DATAD$ is DEV1_ on my machine.

Marcel

___
QL-Users Mailing List


___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Wolf]

Hi,
I often use something like this for ftest

100 DEFine FuNction ftest2(file$)
110 LOCal dad$,ded$,prd$,a
120   dad$=DATAD$
130   ded$=DESTD$
140   prd$=PROGD$
150   DATA_USE ""
160   DEST_USE ""
170   PROG_USE ""
180   a=FTEST(file$)
190   DATA_USE dad$
200   DEST_USE ded$
210   PROG_USE prd$
220   RETurn a
230 END DEFine ftest2
240 :

HTH

Wolfgang
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Norman Dunbar]

Yes.

Cheers,
Norm.

On 14 June 2016 20:53:15 BST, "Ralf Reköndt"  wrote:
>Does this ommit TK2?
>
>- Original Message - 
>From: "George Gwilt"
>
>> Using Turbo Toolkit’s DEVICE_STATUS might be safer.
>>
>> George 
>
>___
>QL-Users Mailing List

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Norman Dunbar]

You could use max_devs and dev_name from the truly wonderful DJTOOLKIT too, if 
you were inclined! I'm surprised  that DJ didn't mention it! ;)

Cheers,
Norm.
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Ralf Reköndt]

Does this ommit TK2?

- Original Message - 
From: "George Gwilt"



Using Turbo Toolkit’s DEVICE_STATUS might be safer.

George 


___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Ralf Reköndt]

Well, in the first place, I always thought, that FTEST makes an direct check 
on the file or device, omitting every TK2 occurence. Unfortunately, I was 
wrong.


- Original Message - 
From: "Dilwyn Jones" <dil...@evans1511.fsnet.co.uk>

To: <ql-us...@q-v-d.com>
Sent: Tuesday, June 14, 2016 6:07 PM
Subject: Re: [Ql-Users] A feature or a bug?



Can someone explain this?
10 f$="":REMark empty string
20 if FTEST(f$)=0
30 DIR f$
40 ELSE
50 PRINT f$," can't be a file name"
60 ENDIF
Why is line 30 executed (displaying the directory of datad$) instead of 
line 50?
IF line 30 is replace with VIEW f$ then the content of win (in my case 
Win2_) is displayed.

Can someone confirm the above program?

François Van Emelen
FTEST("") is essentially FTEST(DATAD$) as Marcel and Tim have pointed out. 
The TK2 extensions like this all pick up the system defaults.


a root drive name such as WIN1_ or WIN2_ on a level 2 system contains the 
list of files details for the directory, as you'll see if you COPY WIN1_ 
TO SCR.


So, I would politely suggest you do NOT put DELETE or WDEL in line 30, 
François!!! (unless you enjoy putting your win drives at risk of 
self-destruct)


I daresay you could use a clumsy coding liek this to work around it:

5 d$=DATAD$:DATA_USE ""
10 f$=""
20 IF FTEST(f$)=0
...
70 DATA_USE d$

(even then there is a the risk another job might alter DATA_USE in between 
lines of your program, just to screw things up, and beware of when DATAD$ 
returns just '_')


As an aside, the null filename used to be used (on level 1 filing systems) 
in protection systems as a kind of hidden file in the early days of QL 
software.


Dilwyn




-
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7639 / Virus Database: 4604/12419 - Release Date: 06/14/16

___
QL-Users Mailing List 


___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Ralf Reköndt]

- Original Message - 
From: "Marcel Kilgus"



Ralf Reköndt wrote:

It's really that awful, as OPEN tries the device you told it, then, if no
success, uses DATAD$ or somewhat others, DEV or such.


DATAD$, PROGD$ or DESTD$ is prepended to the filename, depending on
the call. DEV is just a device like any other I think.

Note that I do not endorse this design in any way, the whole filename
handling in QDOSMSQ is not, let's say, optimal. But as I said it's
been this way for 30 years, I don't see any urgency to change it now.


Yes, but I found it the only way in S_Edit to see, if a filename really 
exists, not to pass it to DATD$ or similar. I set the SV_DATAD to "", kept 
the value, check the device and put it back. There were no other way. Ok, as 
far as I know.


Cheers...Ralf 


___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Marcel Kilgus]

Dilwyn Jones wrote:
> There are device open functions out there in other toolkits which do the
> same job without checking the TK2 defaults system.

This is an important point. This is not an OS feature, i.e. the raw
TRAPs do not behave this way, it's strictly a feature of the BASIC
commands. That was the only way it could be implemented back then as
the OS was fixed. If another utility wraps the open call it will get
the "raw" behaviour.

Ralf Reköndt wrote:
> It's really that awful, as OPEN tries the device you told it, then, if no
> success, uses DATAD$ or somewhat others, DEV or such.

DATAD$, PROGD$ or DESTD$ is prepended to the filename, depending on
the call. DEV is just a device like any other I think.

Note that I do not endorse this design in any way, the whole filename
handling in QDOSMSQ is not, let's say, optimal. But as I said it's
been this way for 30 years, I don't see any urgency to change it now.

All the best, Marcel

___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Dilwyn Jones]

There are device open functions out there in other toolkits which do the 
same job without checking the TK2 defaults system.


Also, it's easy enough to do an error-trapped open_in on most systems, 
whether you use WHEN ERRor (JS or later ROM), Q_ERR_ON (Qliberator 
compiler), WHEN_ERROR and DEVICE_STATUS (Turbo compiler) just to name a few.


Dilwyn

-Original Message- 
From: Ralf Reköndt

Sent: Tuesday, June 14, 2016 7:20 PM
To: ql-us...@q-v-d.com
Subject: Re: [Ql-Users] A feature or a bug?

It's really that awful, as OPEN tries the device you told it, then, if no
success, uses DATAD$ or somewhat others, DEV or such.

To be really successfull, you should set the SV_DATAD to "", keep the value
then try to open the device. This is the only way to see, if *that* device
is not present.

- Original Message - 
From: "Marcel Kilgus" <ql-us...@mail.kilgus.net>

To: <ql-us...@q-v-d.com>
Sent: Tuesday, June 14, 2016 7:57 PM
Subject: Re: [Ql-Users] A feature or a bug?



François Van Emelen wrote:

So, this seems to be a feature... a very dangerous one : replacing DIR
f$ in line 30 with DELETE f$ would wipe all directories and files on
that device.


That would actually be a somewhat cool feature, but DELETE is not as
cool, so nothing will happen. Try it :) WDEL is a little bit more
dangerous, but it still asks nicely and like DELETE it will also fail
on directories.

Thing is this: FTEST = OPEN:CLOSE. OPEN can be used without a
parameter, in which case DATAD$ will be used. The same is therefore
true for FTEST. It's actually been this way since the release of TK2
which has been how long ago, 30 years or so?

All the best, Marcel


___
QL-Users Mailing List

-
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7639 / Virus Database: 4604/12419 - Release Date: 06/14/16 




-
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7639 / Virus Database: 4604/12419 - Release Date: 06/14/16

___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Ralf Reköndt]

It's really that awful, as OPEN tries the device you told it, then, if no 
success, uses DATAD$ or somewhat others, DEV or such.


To be really successfull, you should set the SV_DATAD to "", keep the value 
then try to open the device. This is the only way to see, if *that* device 
is not present.


- Original Message - 
From: "Marcel Kilgus" <ql-us...@mail.kilgus.net>

To: <ql-us...@q-v-d.com>
Sent: Tuesday, June 14, 2016 7:57 PM
Subject: Re: [Ql-Users] A feature or a bug?



François Van Emelen wrote:

So, this seems to be a feature... a very dangerous one : replacing DIR
f$ in line 30 with DELETE f$ would wipe all directories and files on
that device.


That would actually be a somewhat cool feature, but DELETE is not as
cool, so nothing will happen. Try it :) WDEL is a little bit more
dangerous, but it still asks nicely and like DELETE it will also fail
on directories.

Thing is this: FTEST = OPEN:CLOSE. OPEN can be used without a
parameter, in which case DATAD$ will be used. The same is therefore
true for FTEST. It's actually been this way since the release of TK2
which has been how long ago, 30 years or so?

All the best, Marcel 


___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Marcel Kilgus]

François Van Emelen wrote:
> So, this seems to be a feature... a very dangerous one : replacing DIR
> f$ in line 30 with DELETE f$ would wipe all directories and files on 
> that device.

That would actually be a somewhat cool feature, but DELETE is not as
cool, so nothing will happen. Try it :) WDEL is a little bit more
dangerous, but it still asks nicely and like DELETE it will also fail
on directories.

Thing is this: FTEST = OPEN:CLOSE. OPEN can be used without a
parameter, in which case DATAD$ will be used. The same is therefore
true for FTEST. It's actually been this way since the release of TK2
which has been how long ago, 30 years or so?

All the best, Marcel


___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Dilwyn Jones]

I think it's just that Marcel normally just uses DATA_USE dev1_ on his 
system.

Dilwyn

-Original Message- 
From: Derek Stewart

Sent: Tuesday, June 14, 2016 6:27 PM
To: ql-us...@q-v-d.com
Subject: Re: [Ql-Users] A feature or a bug?


On 14/06/16 14:43, Marcel Kilgus wrote:

Marcel Kilgus wrote:

François Van Emelen wrote:

10 f$="":REMark empty string
20 if FTEST(f$)=0
30 DIR f$
40 ELSE
50 PRINT f$," can't be a file name"
60 ENDIF

Why is line 30 executed (displaying the directory of datad$) instead of
line 50?

Because SMSQ/E defaults to dev1_ if you don't give a filename, I
presume.


Strike that, it defaults to datad$ of course, which is set to "dev1_"
in my case. I mean, that's what DATA_USE is for, providing a default
directory for file operations.

Marcel

___
QL-Users Mailing List


Hi Marcel,

I am a little confused, where does DEV1_ default to DATAD$

If I startup QPC2 with no boot device defined, the DATAD$ is flp1_,
which is the default of config block in the SMSQE file and DEV1_ is
undefined.

Can expand on this, I am havong a thick moment...


Regards,

Derek
___
QL-Users Mailing List

-
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7639 / Virus Database: 4604/12419 - Release Date: 06/14/16 




-
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7639 / Virus Database: 4604/12419 - Release Date: 06/14/16

___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Marcel Kilgus]

Derek Stewart wrote:
> I am a little confused, where does DEV1_ default to DATAD$

Never. I said DATAD$ is DEV1_ on my machine.

Marcel

___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Derek Stewart]

On 14/06/16 14:43, Marcel Kilgus wrote:

Marcel Kilgus wrote:

François Van Emelen wrote:

10 f$="":REMark empty string
20 if FTEST(f$)=0
30 DIR f$
40 ELSE
50 PRINT f$," can't be a file name"
60 ENDIF

Why is line 30 executed (displaying the directory of datad$) instead of
line 50?

Because SMSQ/E defaults to dev1_ if you don't give a filename, I
presume.


Strike that, it defaults to datad$ of course, which is set to "dev1_"
in my case. I mean, that's what DATA_USE is for, providing a default
directory for file operations.

Marcel

___
QL-Users Mailing List


Hi Marcel,

I am a little confused, where does DEV1_ default to DATAD$

If I startup QPC2 with no boot device defined, the DATAD$ is flp1_, 
which is the default of config block in the SMSQE file and DEV1_ is 
undefined.


Can expand on this, I am havong a thick moment...


Regards,

Derek
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[George Gwilt]

> On 14 Jun 2016, at 16:45, François Van Emelen  
> wrote:
> 
> Op 14/06/2016 om 16:18 schreef George Gwilt:
>>> On 14 Jun 2016, at 14:56, Bob Spelten  wrote:
>>> 
>>> Op Tue, 14 Jun 2016 15:43:00 +0200 schreef Marcel Kilgus 
>>> :
>>> 
 Marcel Kilgus wrote:
> François Van Emelen wrote:
>> 10 f$="":REMark empty string
>> 20 if FTEST(f$)=0
>> 30 DIR f$
>> 40 ELSE
>> 50 PRINT f$," can't be a file name"
>> 60 ENDIF
>> 
>> Why is line 30 executed (displaying the directory of datad$) instead of
>> line 50?
> Because SMSQ/E defaults to dev1_ if you don't give a filename, I
> presume.
 Strike that, it defaults to datad$ of course, which is set to "dev1_"
 in my case. I mean, that's what DATA_USE is for, providing a default
 directory for file operations.
 
 Marcel
 
>>> I understand the datad$ but the problem is why FTEST returns a zero?
>>> FTEST should return a -7, so DIR is should not be done.
>>> 
>>> 
>> The manual says that FTEST tests "a file or device”. Thus FTESTing win4_, 
>> say, returns “0” even if win4_ contains no files.
>> 
>> George
>> 
>> ___
>> QL-Users Mailing List
> 
> So, this seems to be a feature... a very dangerous one : replacing DIR f$ in 
> line 30 with DELETE f$ would wipe all directories and files on that device.
> 
> And there is no Undo available .
> 
> François Van Emelen
> 
> 
> 
> ___
> QL-Users Mailing List

Using Turbo Toolkit’s DEVICE_STATUS might be safer.

George
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Dilwyn Jones]

Can someone explain this?
10 f$="":REMark empty string
20 if FTEST(f$)=0
30 DIR f$
40 ELSE
50 PRINT f$," can't be a file name"
60 ENDIF
Why is line 30 executed (displaying the directory of datad$) instead of 
line 50?
IF line 30 is replace with VIEW f$ then the content of win (in my case 
Win2_) is displayed.

Can someone confirm the above program?

François Van Emelen
FTEST("") is essentially FTEST(DATAD$) as Marcel and Tim have pointed out. 
The TK2 extensions like this all pick up the system defaults.


a root drive name such as WIN1_ or WIN2_ on a level 2 system contains the 
list of files details for the directory, as you'll see if you COPY WIN1_ TO 
SCR.


So, I would politely suggest you do NOT put DELETE or WDEL in line 30, 
François!!! (unless you enjoy putting your win drives at risk of 
self-destruct)


I daresay you could use a clumsy coding liek this to work around it:

5 d$=DATAD$:DATA_USE ""
10 f$=""
20 IF FTEST(f$)=0
...
70 DATA_USE d$

(even then there is a the risk another job might alter DATA_USE in between 
lines of your program, just to screw things up, and beware of when DATAD$ 
returns just '_')


As an aside, the null filename used to be used (on level 1 filing systems) 
in protection systems as a kind of hidden file in the early days of QL 
software.


Dilwyn




-
No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7639 / Virus Database: 4604/12419 - Release Date: 06/14/16

___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[François Van Emelen]

Op 14/06/2016 om 16:18 schreef George Gwilt:

On 14 Jun 2016, at 14:56, Bob Spelten  wrote:

Op Tue, 14 Jun 2016 15:43:00 +0200 schreef Marcel Kilgus 
:


Marcel Kilgus wrote:

François Van Emelen wrote:

10 f$="":REMark empty string
20 if FTEST(f$)=0
30 DIR f$
40 ELSE
50 PRINT f$," can't be a file name"
60 ENDIF

Why is line 30 executed (displaying the directory of datad$) instead of
line 50?

Because SMSQ/E defaults to dev1_ if you don't give a filename, I
presume.

Strike that, it defaults to datad$ of course, which is set to "dev1_"
in my case. I mean, that's what DATA_USE is for, providing a default
directory for file operations.

Marcel


I understand the datad$ but the problem is why FTEST returns a zero?
FTEST should return a -7, so DIR is should not be done.



The manual says that FTEST tests "a file or device”. Thus FTESTing win4_, say, 
returns “0” even if win4_ contains no files.

George

___
QL-Users Mailing List


So, this seems to be a feature... a very dangerous one : replacing DIR 
f$ in line 30 with DELETE f$ would wipe all directories and files on 
that device.


And there is no Undo available .

François Van Emelen



___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[George Gwilt]

> On 14 Jun 2016, at 14:56, Bob Spelten  wrote:
> 
> Op Tue, 14 Jun 2016 15:43:00 +0200 schreef Marcel Kilgus 
> :
> 
>> Marcel Kilgus wrote:
>>> François Van Emelen wrote:
 10 f$="":REMark empty string
 20 if FTEST(f$)=0
 30 DIR f$
 40 ELSE
 50 PRINT f$," can't be a file name"
 60 ENDIF
 
 Why is line 30 executed (displaying the directory of datad$) instead of
 line 50?
>>> Because SMSQ/E defaults to dev1_ if you don't give a filename, I
>>> presume.
>> 
>> Strike that, it defaults to datad$ of course, which is set to "dev1_"
>> in my case. I mean, that's what DATA_USE is for, providing a default
>> directory for file operations.
>> 
>> Marcel
>> 
> I understand the datad$ but the problem is why FTEST returns a zero?
> FTEST should return a -7, so DIR is should not be done.
> 
> 

The manual says that FTEST tests "a file or device”. Thus FTESTing win4_, say, 
returns “0” even if win4_ contains no files.

George

___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Marcel Kilgus]

Bob Spelten wrote:
> I understand the datad$ but the problem is why FTEST returns a zero?

I don't understand, why shouldn't it?

> FTEST should return a -7, so DIR is should not be done.

Clearly the directory pointed to by DATAD$ exists, so why whould it
return -7?

Marcel



___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Bob Spelten]

Op Tue, 14 Jun 2016 15:43:00 +0200 schreef Marcel Kilgus  
:



Marcel Kilgus wrote:

François Van Emelen wrote:

10 f$="":REMark empty string
20 if FTEST(f$)=0
30 DIR f$
40 ELSE
50 PRINT f$," can't be a file name"
60 ENDIF

Why is line 30 executed (displaying the directory of datad$) instead of
line 50?

Because SMSQ/E defaults to dev1_ if you don't give a filename, I
presume.


Strike that, it defaults to datad$ of course, which is set to "dev1_"
in my case. I mean, that's what DATA_USE is for, providing a default
directory for file operations.

Marcel


I understand the datad$ but the problem is why FTEST returns a zero?
FTEST should return a -7, so DIR is should not be done.

Bob

--
The BSJR QL software site at: http://members.upc.nl/b.spelten/ql/
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Timothy Swenson]

I tried this and it worked as reported.  I then did this:

data_use ""

then I ran it again, and got the error message.  This confirms that 
datad$ is being picked up and used in the FTEST.


Tim
___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Marcel Kilgus]

Marcel Kilgus wrote:
> François Van Emelen wrote:
>> 10 f$="":REMark empty string
>> 20 if FTEST(f$)=0
>> 30 DIR f$
>> 40 ELSE
>> 50 PRINT f$," can't be a file name"
>> 60 ENDIF
>>
>> Why is line 30 executed (displaying the directory of datad$) instead of
>> line 50?
> Because SMSQ/E defaults to dev1_ if you don't give a filename, I
> presume.

Strike that, it defaults to datad$ of course, which is set to "dev1_"
in my case. I mean, that's what DATA_USE is for, providing a default
directory for file operations.

Marcel

___
QL-Users Mailing List

Re: [Ql-Users] A feature or a bug?

[Marcel Kilgus]

François Van Emelen wrote:
> 10 f$="":REMark empty string
> 20 if FTEST(f$)=0
> 30 DIR f$
> 40 ELSE
> 50 PRINT f$," can't be a file name"
> 60 ENDIF
>
> Why is line 30 executed (displaying the directory of datad$) instead of
> line 50?

Because SMSQ/E defaults to dev1_ if you don't give a filename, I
presume.

Marcel

___
QL-Users Mailing List

[Ql-Users] A feature or a bug?

[François Van Emelen]

Hi,

Can someone explain this?

10 f$="":REMark empty string

20 if FTEST(f$)=0

30 DIR f$

40 ELSE

50 PRINT f$," can't be a file name"

60 ENDIF

Why is line 30 executed (displaying the directory of datad$) instead of 
line 50?


IF line 30 is replace with VIEW f$ then the content of win (in my case 
Win2_) is displayed.


Can someone confirm the above program?

François Van Emelen



___
QL-Users Mailing List